Lowest SLP, Maximum Sustained Winds and Gusts

# subset section A
sec_a <- txt[grep("^A\\. LOWEST SEA LEVEL PRESSURE", txt):(grep("^B\\. MARINE OBSERVATIONS", txt) - 1)]
# trim head
sec_a <- sec_a[-c(1:9)]
# trim tail
sec_a <- head(sec_a, n = -4L)

# subset section B
sec_b <- txt[grep("^B\\. MARINE OBSERVATIONS", txt):(grep("^C\\. STORM TOTAL RAINFALL", txt) - 1)]
# trim head
sec_b <- sec_b[-c(1:8)]
# trim tail
sec_b <- head(sec_b, n = -5L)

# Bind to sec
sec <- c(sec_a, sec_b)

# There is a subheader in middle I'll also manually remove
sec <- sec[-c(78:88)]

# Get station header data
stations <- str_match(sec, 
                      sprintf("^%s[\\s-]*?%s[\\s-]*?%s[\\s]+$", 
                              "([[K|X|J|4]\\w]{3,5})",   # Station
                              "([\\w\\s/\\.\\(\\)]+)",       # Location
                              "([\\D]{2})")) %>%             # State
  .[complete.cases(.),]

# Get station obs. I really wanted to expand the regex to do this; use it as 
# a learning opportunity. But, apparently my skills aren't quite that 
# advanced... oh, well
obs <- str_match(sec, "^(\\d{2}\\.\\d{2}.+$)")[,2] %>% .[complete.cases(.)]

# Move to dataframe
slp <- as_data_frame(
  list("Station" = stations[,2], 
       "Location" = stations[,3], 
       "State" = stations[,4], 
       "Lat" = as.numeric(str_sub(obs, 0L, 5L)), 
       "Lon" = as.numeric(str_sub(obs, 7L, 12L)), 
       "MinPres" = as.numeric(str_sub(obs, 16L, 21L)), 
       "MinPresDt" = as.POSIXct(ymd_hm(sprintf("2017-08-%s %s", 
                                               str_sub(obs, 23L, 24L), 
                                               str_sub(obs, 26L, 29L))), 
                               tz = "UTC"), 
       "Remarks" = str_sub(obs, 31L, 31L), 
       "MaxSustDir" = as.numeric(str_sub(obs, 33L, 35L)), 
       "MaxSustSpd" = as.numeric(str_sub(obs, 37L, 39L)),
       "MaxSustDt" = as.POSIXct(ymd_hm(sprintf("2017-08-%s %s", 
                                               str_sub(obs, 42L, 43L), 
                                               str_sub(obs, 45L, 48L))), 
                               tz = "UTC"), 
       "PeakGustDir" = as.numeric(str_sub(obs, 52L, 54L)), 
       "PeakGustSpd" = as.numeric(str_sub(obs, 56L, 58L)),
       "PeakGustDt" = as.POSIXct(ymd_hm(sprintf("2017-08-%s %s", 
                                                str_sub(obs, 60L, 61L), 
                                                str_sub(obs, 63L, 66L))), 
                               tz = "UTC"))) %>% 
  mutate(Lat = if_else(Lat < 0, Lat * -1, Lat), 
         Lon = if_else(Lon > 0, Lon * -1, Lon))

## Warning: 13 failed to parse.

## Warning: 1 failed to parse.

## Warning: 1 failed to parse.

# Clean up
slp$MinPres[slp$MinPres == 9999.0] <- NA
slp$MaxSustDir[slp$MaxSustDir == 999] <- NA
slp$MaxSustSpd[slp$MaxSustSpd == 999] <- NA

p <- bp + 
  geom_point(data = slp, 
             aes(x = Lon, y = Lat, 
                 text = sprintf("%s\n%s, %s\n%s UTC", 
                                Station, 
                                Location, 
                                State, 
                                MaxSustDt), 
                 size = MaxSustSpd, color = MaxSustSpd)) + 
  coord_equal(xlim = c(min(slp$Lon, na.rm = TRUE), max(slp$Lon, na.rm = TRUE)), 
              ylim = c(min(slp$Lat, na.rm = TRUE), max(slp$Lat, na.rm = TRUE)))

## Warning: Ignoring unknown aesthetics: text

ggplotly(p, tooltip = c("text", "size", "color"))

Storm Total Rainfall

Storm total rainfall from Aug. 25, 12:00 UTC (7:00 AM CDT) through Aug. 31 12:00 UTC (7:00 AM CDT)

rain_obs <- txt[grep("^C\\. STORM TOTAL RAINFALL", txt):(grep("^D\\. INLAND FLOODING", txt) - 1)]

# Drop header and tail data
rain_obs <- rain_obs[-c(1:6)]
rain_obs <- head(rain_obs, -2L)

# Drop empties
rain_obs <- rain_obs[rain_obs != ""]

# Here I'll get observations first
pos <- str_match(rain_obs, sprintf("^%s[\\s]+%s[\\s]?$", 
                                 "([\\d\\.]{5})", 
                                 "([-\\d\\.]{6})"))
# Remove NAs
pos <- pos[complete.cases(pos),]

# Get location details
loc <- str_match(rain_obs, sprintf("^%s[\\s]{3,}%s[\\s]{3,}%s[\\s]{3,}%s[\\s]+%s$", 
                                   "([\\w\\d\\s\\.\\-/]{1,29})", # Location
                                   "([\\w\\s]{1,20})",           # County
                                   "([\\w\\d-=]*)",         # ID
                                   "([\\d\\.]{3,5})",           # Rainfall
                                   "(I*)")) %>%             # Remarks
  .[complete.cases(.),]

# Make df
rain <- as_data_frame(
  list("Location" = loc[,2], 
       "County" = loc[,3], 
       "ID" = loc[,4], 
       "Rainfall" = as.numeric(loc[,5]), 
       "Remarks" = loc[,6],
       "Lat" = as.numeric(pos[,2]), 
       "Lon" = as.numeric(pos[,3]))) %>% 
  mutate(Lat = if_else(Lat < 0, Lat * -1, Lat), 
         Lon = if_else(Lon > 0, Lon * -1, Lon))

p <- bp + 
  geom_point(data = rain, 
             aes(x = Lon, y = Lat, 
                 text = sprintf("%s\n%s, %s", 
                                ID, 
                                Location, 
                                County), 
                 size = Rainfall, color = Rainfall)) + 
  coord_equal(xlim = c(min(rain$Lon, na.rm = TRUE), max(rain$Lon, na.rm = TRUE)), 
              ylim = c(min(rain$Lat, na.rm = TRUE), max(rain$Lat, na.rm = TRUE)))

## Warning: Ignoring unknown aesthetics: text

ggplotly(p, tooltip = c("text", "size", "color"))

## We recommend that you use the dev version of ggplot2 with `ggplotly()`
## Install it with: `devtools::install_github('hadley/ggplot2')`

Maximum Storm Surge and Storm Tide

tide_obs <- txt[grep("^E\\. MAXIMUM STORM SURGE AND STORM TIDE", txt):(grep("^F\\. TORNADOES", txt) - 1)]
# Trim the edges
tide_obs <- tide_obs[-c(1:7)]
tide_obs <- head(tide_obs, n = -5L)
tide_obs[tide_obs == ""] <- NA
tide_obs <- tide_obs[complete.cases(tide_obs)]

tide <- as_data_frame(
  list("County" = str_trim(str_sub(tide_obs, 0L, 16L)), 
       "Location" = str_trim(str_sub(tide_obs, 18L, 32L)), 
       "Surge" = as.numeric(str_trim(str_sub(tide_obs, 34L, 39L))), 
       "Tide" = as.numeric(str_trim(str_sub(tide_obs, 42L, 46L))), 
       "DateTime" = as.POSIXct(ymd_hm(sprintf("2017-08-%s %s", 
                                              str_sub(tide_obs, 49L, 50L), 
                                              str_sub(tide_obs, 52L, 55L))), 
                               tz = "UTC"), 
       "Erosion" = str_trim(str_sub(tide_obs, 58L, 64L))))

kable(tide, caption = "Storm Surge and Storm Tide Totals (ft)")

Tornadoes

tor_obs <- txt[grep("^F\\. TORNADOES", txt):(grep("^G\\. STORM IMPACTS BY COUNTY", txt) - 1)]
tor_obs <- tor_obs[-c(1:6)]
tor_obs <- head(tor_obs, n = -3L)

# county, time and tornado scale
loc <- str_match(tor_obs, "^([\\w\\d\\s]{28})\\s+([\\w\\s]{16})\\s+([\\d/]{7})\\s+(EF\\d)\\s+$") %>% 
  .[complete.cases(.),]

# lat, lon
pos <- str_match(tor_obs, "^([\\d\\.]{4,5})\\s+([-\\d\\.]{4,6})$") %>% 
  .[complete.cases(.),]

# Now, get the remarks. Subtract loc and pos to make it easier.
# Essentially, what I'm doing here is joining the vector together 
# with a \t. Consecutive \t's indicate a break between remarks so 
# make that a newline. Then split on \n and trim, return to vector.
rmks <- tor_obs[!(tor_obs %in% c(loc, pos))]
rmks <- str_c(rmks, collapse = "\t")
rmks <- str_replace_all(rmks, "\t\t+", "\n")
rmks <- str_replace_all(rmks, "\t", " ")
rmks <- str_split(rmks, "\n")
rmks <- map(rmks, str_trim, side = "both") %>% flatten_chr()

tors <- as_data_frame(
  list("Location" = loc[,2], 
       "County" = loc[,3], 
       "Date" = as.POSIXct(ymd_hm(sprintf("2017-08-%s %s", 
                                          str_sub(loc[,4], 0L, 2L), 
                                          str_sub(loc[,4], 4L, 7L)))), 
       "Scale" = factor(loc[,5], 
                        levels = c("EF0", "EF1", "EF2", "EF3", "EF4", "EF5"), 
                        labels = c("EF0", "EF1", "EF2", "EF3", "EF4", "EF5")), 
       "Lat" = as.numeric(pos[,2]), 
       "Lon" = as.numeric(pos[,3]), 
       "Remarks" = rmks))

tmp <- tors %>% filter(Lat > 0, Lon < 0) # Filter out missing values
p <- bp + 
  geom_point(data = tmp, 
             aes(x = Lon, y = Lat, 
                 text = sprintf("%s", 
                                str_wrap(Remarks)), 
                 size = Scale, color = Scale)) + 
  coord_equal(xlim = c(min(tmp$Lon, na.rm = TRUE), max(tmp$Lon, na.rm = TRUE)), 
              ylim = c(min(tmp$Lat, na.rm = TRUE), max(tmp$Lat, na.rm = TRUE)))

## Warning: Ignoring unknown aesthetics: text

ggplotly(p, tooltip = c("text", "size", "color"))

## We recommend that you use the dev version of ggplot2 with `ggplotly()`
## Install it with: `devtools::install_github('hadley/ggplot2')`

## Warning: Using size for a discrete variable is not advised.

kable(tors, caption = "Tornadoes")

Source

Post Tropical Cyclone Report - Tropical Storm Harvey

Session Info

pander::pander(sessionInfo())

R version 3.4.0 (2017-04-21)

**Platform:** x86_64-pc-linux-gnu (64-bit)

locale: LC_CTYPE=en_US.UTF-8, LC_NUMERIC=C, LC_TIME=en_US.UTF-8, LC_COLLATE=en_US.UTF-8, LC_MONETARY=en_US.UTF-8, LC_MESSAGES=C, LC_PAPER=en_US.UTF-8, LC_NAME=C, LC_ADDRESS=C, LC_TELEPHONE=C, LC_MEASUREMENT=en_US.UTF-8 and LC_IDENTIFICATION=C

attached base packages: methods, stats, graphics, grDevices, utils, datasets and base

other attached packages: bindrcpp(v.0.2), tidyr(v.0.6.3), tibble(v.1.3.3), stringr(v.1.2.0), rrricanes(v.0.2.0-6), purrr(v.0.2.2.2), plotly(v.4.7.1), lubridate(v.1.6.0), knitr(v.1.16), ggrepel(v.0.6.5), ggplot2(v.2.2.1) and dplyr(v.0.7.2)

loaded via a namespace (and not attached): Rcpp(v.0.12.12), rrricanesdata(v.0.0.1.4), highr(v.0.6), compiler(v.3.4.0), plyr(v.1.8.4), bindr(v.0.1), tools(v.3.4.0), digest(v.0.6.12), lattice(v.0.20-35), viridisLite(v.0.2.0), jsonlite(v.1.5), evaluate(v.0.10.1), gtable(v.0.2.0), pkgconfig(v.2.0.1), rlang(v.0.1.1), shiny(v.1.0.3), crosstalk(v.1.0.0), yaml(v.2.1.14), blogdown(v.0.0.56), rnaturalearthdata(v.0.1.0), httr(v.1.2.1), htmlwidgets(v.0.9), rprojroot(v.1.2), grid(v.3.4.0), glue(v.1.1.1), data.table(v.1.10.4), R6(v.2.2.2), rmarkdown(v.1.6), bookdown(v.0.4), sp(v.1.2-5), pander(v.0.6.0), magrittr(v.1.5), backports(v.1.1.0), scales(v.0.4.1), htmltools(v.0.3.6), assertthat(v.0.2.0), xtable(v.1.8-2), mime(v.0.5), colorspace(v.1.3-2), httpuv(v.1.3.5), labeling(v.0.3), stringi(v.1.1.5), lazyeval(v.0.2.0) and munsell(v.0.4.3)